Frequently Asked Questions

General

What is the purpose of this tool?

ThinkHazard! was designed to increase awareness of natural hazards anywhere in the world. It does so by providing a simple overview of different hazards and their ‘hazard levels’ for any given area, delivering recommendations and information to aid users in managing their risk to those hazards. The tool is designed to become increasingly comprehensive over time as users contribute new data and information. If you have data or information to contribute, please click here. ThinkHazard! does not provide tools for detailed risk analysis. IDF CatRiskTools provides a catalog of open access and commercially licenced risk analysis tools. Impact analysis can be conducted on data stored at GFDRR Innovation Lab GeoNode using the new GeoSafe impact analysis tool.



What do the hazard levels mean?

The classification of hazard is based on the likelihood of the hazard exceeding predefined thresholds. The thresholds are based on hazard frequency and intensity and set using expert judgement. A higher hazard classification in ThinkHazard! indicates that there is greater potential for damage or disruption to activities or a project in that region, according to the underlying hazard data.



Does ThinkHazard! provide the hazard level at my specific location of interest?

No. ThinkHazard! provides the maximum hazard level within each district, province, or national administrative unit (administrative level 2, 1, and 0, respectively). It does not provide the hazard level at specific locations. This tool is a starting point to increase awareness of the hazards present in an area. To determine the hazard level at a specific location, users should refer to more detailed hazard maps for the area, for example the underlying hazard data provided in the GFDRR Innovation Lab GeoNode or through contacting the relevant local authorities.



How does ThinkHazard! work?

ThinkHazard! uses the best available hazard data as the basis of its classification of hazard level. One method uses probabilistic data, which specifies the intensity and frequency of events in each region – e.g., earthquake data show the likelihood of a particular amount of ground shaking occurring at different locations. ThinkHazard! also uses data that says, for example, ‘the climate and ground conditions in this region suggest the region has a certain susceptibility to landslides’.

The classification of hazard level in ThinkHazard! is based on the likelihood of each region experiencing a hazard with enough intensity (e.g., intensity of ground shaking) to cause damage. A higher hazard classification in ThinkHazard! indicates that there is greater potential for damage or disruption to activities or a project in that region, according to the underlying hazard data.

For any region selected by a user, ThinkHazard! shows the maximum hazard level within that region. In the example below, earthquake hazard in the district ‘Ladislao Cabrera’ is rated ‘Medium’ based on underlying hazard data, but some of its neighboring districts are rated ‘Low’ (left image). The parent region, ‘Oruro’ Departmento (right image), is assigned the maximum hazard level of all districts it contains, so it is also assigned ‘Medium’. Bolivia contains nine Departmentos, which have different hazard levels, but as maximum hazard in ‘Medium’, Bolivia is assigned ‘Medium’. This example demonstrates the need to drill down to the region(s) relevant to your project, to find out more about the distribution of hazard there.

Oruro

Bolivia



Does ThinkHazard! use real-time data and historical event databases?

We do not provide real-time event information or historical databases of events in ThinkHazard!. Websites that provide historical event databases or information on current and recent events are provided here.



Has the information in ThinkHazard! been peer-reviewed?

We commissioned an independent analysis of how well ThinkHazard! estimated hazard levels for up to 77 developing countries around the world and the results can be seen here. We recognize the importance of producing accurate results for our users so we are continuously looking for more accurate national and local datasets to supplement the tool. If you are aware of additional hazard datasets please contact us via the feedback form, available on every page.



Disasters don’t stop at borders or administrative boundaries. Why is ThinkHazard! based on (national, provincial, district) boundaries?

While we understand that presenting hazard data by administrative boundaries might not provide a full view of hazard data, it is the simplest way to visualize complex hazard data. The underlying hazard data have different levels of resolution, and presentation by boundary provides the most consistent output. The underlying grid data, which is not shown at administration boundary level, can be accessed at the GFDRR Innovation Lab GeoNode.



Are there more hazards in my area than are shown?

There may be hazards that affect your area that are not shown in ThinkHazard!. The tool focuses on natural hazards only, so hazards such as road traffic hazards and chemical spill hazards are not considered. Localized natural hazards such as avalanche and tornado, or those hazards that are specific to very few countries, are not presented. If there are hazards that you feel should be presented, please record this using the feedback form, available on every page.



A disaster recently occurred in a region with low or medium hazard. Does this mean the hazard level is wrong?

Not necessarily. The hazard level shown in ThinkHazard! is based on the expected frequency of the hazard exceeding intensity thresholds over the long-term (decades to centuries). If the hazard level is low, it means that the hazard is unlikely to exceed the threshold in that region, in that timeframe. The fact that an event occurred recently does not necessarily change the average frequency in the long-term.

Similarly, the hazard level may be considered high, even if there has not been an event in living memory. Some hazards may only occur years or decades apart in some regions, but are classified as high hazard when looking at longer timescales. These are known as ‘low frequency hazards’, and include earthquakes, volcanic eruptions, and tsunamis.



What does return period mean?

A return period expresses the likelihood that an event could happen in a given year. The larger the return period for an event, the less likely its occurrence, and the greater its intensity. Return period, expresses the average time between hazard events of a particular severity, or intensity.

It is important to recognize that these numbers are averages. When an event is referred to as having a ‘1 in 100-year return period’, it means that on average, an event that size or bigger would be expected to occur once in a 100-year period. However, another event of that size or bigger could practically occur in the next day, or year, or in another 50 years. Or, one that size may not occur again for more than 100 years. Remember, these are statistical averages.

An equivalent way of expressing return period is as the chance of occurring each year. A ‘1 in 100-year return period’ also means a 1% chance of occurring in any single year. A more frequent event, say one with ‘1 in 20-year return period’ has a 5% chance of occurring in any single year (1/20 = 0.05, or 5%).



What is ‘water scarcity’?

Water scarcity is a combination of water availability and water demand. ThinkHazard! incorporates drought through a measure of water scarcity.



About

Is ThinkHazard! open source?

Yes! The code for ThinkHazard! is available on Github. Open source technology means that anyone can adapt the tool for their own use and jointly contribute to the ongoing development of ThinkHazard!.



Is ThinkHazard! available in multiple languages?

The tool is now available in English, French, and Spanish. We will consider additional language functionality in future development. Please let us know what language you would like to see using the feedback form.



I can see a use for this kind of tool in my institution and/or sector. How do I develop a customized version?

We’re glad you asked! Other organizations have already seen a use for this type of tool, from the occupational health and safety sector to disaster risk management professionals working on school safety. The code is open source and is designed to be tailored to different sectors and organizations. Please use the feedback form to let us know how ThinkHazard! could be useful to you so that we can get in touch to discuss development.



What is the long-term vision for the tool?

We want ThinkHazard! to be the go-to source of information for everyone to understand their exposure to natural hazards around the world, to be a reliable source for guidance on how to reduce risk when planning projects, to be a repository of useful reports, websites, and information, and, ultimately, to be a gateway to higher resolution datasets. In order to be a useful source, the data and information that are used in the tool must continually be updated and improved. If you have relevant data or information, please use the feedback form so that we can contact you to discuss.



Improvements and suggestions

Can I provide suggestions to the tool or to the recommendations?

Yes! We continue to improve the tool and want to know what you think. Tell us here.



I have relevant information. Can I contribute this to the tool?

Yes! We are always looking for information and hazard data that is relevant to ThinkHazard! users and encourage you to tell us what you could provide using the feedback form, and we will follow up directly with you. Your institution will be credited as the source of data.



How can I report an error in data or recommendations?

Please let us know about it using the feedback form so that we can try to fix it. Thanks for letting us know, we need your help to constantly improve ThinkHazard!.



Can I get more specific recommendations relevant to my type of project?

This version of ThinkHazard! is designed to be generic, and recommendations are not tailored to any particular sector or project type. If you would like to design a sector- or organization-specific tool (e.g., an education sector tool could provide recommendations on school construction), please contact us using the feedback form to discuss development. The code is open source, available on Github, and is designed to be tailored to sector or organization.



Information for technical users

How does ThinkHazard! work?

To determine the hazard level with probabilistic data, we process the frequency and intensity information contained in each original hazard dataset. First, we set intensity thresholds (using original intensity units), above which the intensity is damaging (not specific to any particular infrastructure or environment).

In some cases, we set a higher threshold for ‘High’ hazard level, than for lower levels, to differentiate the level of damage. We then consider whether the threshold is exceeded at three return periods, which differs for each hazard, reflecting the different frequency of events. If the intensity threshold is exceeded at the most frequent return period, the hazard is determined to be ‘High’. If the threshold is exceeded at the intermediate return period but not the most frequent, it is determined to be ‘Medium’. If the threshold is exceeded at the least frequent return period only, the hazard is determined to be ‘Low’. If there is hazard data present but the threshold is not exceeded at any return period, hazard is determined to be ‘Very Low’.

For non-probabilistic data, such as landslide susceptibility index, we categorize the index into hazard levels using the original data values. The maximum hazard level in each administrative unit 2 (based on the FAO Global Administrative Unit Layers dataset) is deemed to be the hazard level of that unit. The hazard level of a state or province boundary is the maximum hazard level of all districts that it contains. For a full explanation of the methodology, please see the documentation.



Where can I find more information about the methods used in ThinkHazard!?

Please see the documentation.



What hazard data are used in ThinkHazard!, and can I download the data for use elsewhere?

The data sources for each region are listed on the hazard-specific recommendations page, beneath the map. The original data has been processed for use in ThinkHazard! to produce aggregate hazard levels. By clicking on the data source name, you can download any data that are openly available. All data are stored on the GFDRR Innovation Lab GeoNode. Data with commercial limitations on sharing are not available to download. Other open hazard and risk data can be found at sites such Open Data for Resilience Index and OasisHub.



How do you decide which dataset to use if there are multiple datasets for a region?

Where multiple hazard datasets are available for a single region and hazard, ThinkHazard! uses a decision tree to decide which dataset should be used. Submitted hazard data is reviewed by GFDRR hazard and risk specialists. A rating is assigned on the basis of age, methodology used, assumptions, and resolution of the data. Data with a higher rating will be the dataset underlying the results that users see.



The hazard data seem to contradict other data I have seen or used. Why is this?

It is difficult for us to know the cause of the discrepancy without further information. ThinkHazard! uses a variety of hazard data sources. If you have any data that is relevant to hazard levels, please let us know so that we can use it to improve the information on ThinkHazard!.



How does ThinkHazard! account for permanent water bodies in river flood hazard?

ThinkHazard! uses river flood hazard datasets that include a 2-year and/or 5-year return period layer in its calculation of hazard level. One of these layers (2-year if available, otherwise 5-year) are used to exclude areas of permanent water from the calculation. If areas of permanent water were included in the calculation, almost all administrative areas would have ‘High’ hazard. We use the return period layers in place of a global water body mask, because a global mask layer was found to not effectively exclude small rivers. This also resulted in almost all administrative areas being considered as ‘High’ hazard. Datasets without a 2-year and/or 5-year return period layers cannot currently be used in the tool.